Xymon Mailing List Archive search

Hobbit server doesn't start on CentOS 5 with hobbit-4.2.0 installed

list Josh Luthman
Fri, 30 Nov 2007 09:57:57 -0500
Message-Id: <user-0efffd54bba0@xymon.invalid>

Do you have SELinux installed and running by chance?

On 11/30/07, Sebastian <user-7b2156f36779@xymon.invalid> wrote:
 I uninstalled hobbit-4.2.0-7.el5.re.i386.rpm and compiled hobbit-4.2.0from source with the
allinone.patch and get exactly the same error.

It might be worth noting that this was originally a minimum CentOS 5
install - I had to install gcc-c++ (and rrd-devel and pcre-devel) to compile
hobbit, and I previously had to install gcc for compiling some other
software - all from yum.

The server has 2 GB of RAM but is running several other services.

Any other ideas?

Sebastian


*From:* Josh Luthman [mailto:user-4c45a83f15cb@xymon.invalid]
*Sent:* 29 November 2007 18:44
*To:* user-ae9b8668bcde@xymon.invalid
*Subject:* Re: [hobbit] Hobbit server doesn't start on CentOS 5 with
hobbit-4.2.0-7.el5.re.i386.rpm installed

Just compile it from source.  If you don't know how to do that I can give
you instructions command by command, just ask.  I use CentOS5 on my boxes
but I've never bothered with the RPM.

On 11/29/07, Sebastian <user-7b2156f36779@xymon.invalid> wrote:
 I installed an RPM of hobbit-4.2.0 called
hobbit-4.2.0-7.el5.re.i386.rpm, which I probably got from here:
*http://rpm.razorsedge.org/centos-5/RE/repodata/repoview/hobbit-4-4.2.0-7.el5.re.html
*<http://rpm.razorsedge.org/centos-5/RE/repodata/repoview/hobbit-4-4.2.0-7.el5.re.html>;
It includes the All-in-one patch.  So I configured the cfg files and
started hobbit, but I get this in hobbitlaunch.log:

2007-11-29 17:33:47 hobbitlaunch starting
2007-11-29 17:33:47 Loading tasklist configuration from
/usr/libexec/hobbit/server/etc/hobbitlaunch.cfg
2007-11-29 17:33:47 Loading hostnames
2007-11-29 17:33:47 Loading saved state
2007-11-29 17:33:47 Cannot access checkpoint file
/usr/libexec/hobbit/server/tmp/hobbitd.chk for restore
2007-11-29 17:33:47 Setting up network listener on 0.0.0.0:1984
2007-11-29 17:33:47 Setting up signal handlers
2007-11-29 17:33:47 Setting up hobbitd channels
2007-11-29 17:33:47 Could not get shm of size 262144: No space left on
device
2007-11-29 17:33:47 Cannot setup status channel
2007-11-29 17:33:47 Task hobbitd terminated, status 1

This is the first time I've started it, so I presume that this is
normal:  "Cannot access checkpoint file
/usr/libexec/hobbit/server/tmp/hobbitd.chk for restore".  The problem looks
like this:  "Could not get shm of size 262144: No space left on device".  I
see that there are two similar issues listed on *http://www.hswn.dk/hobbit/help/known-issues.html
* <http://www.hswn.dk/hobbit/help/known-issues.html>that relate to
FreeBSD, but none of the sysctl keys listed are in this version of Linux:

# cat /proc/version
Linux version 2.6.18-8.1.15.el5.centos.plus (user-cbb31f498875@xymon.invalid
) (gcc version 4.1.1 20070105 (Red Hat 4.1.1-52)) #1 SMP Thu Oct 25
06:49:40 EDT 2007

Google returns 1 result for "Could not get shm of size" "No space left
on device" and that was the Known Issues page I've already checked.  Any
ideas?

Also, I noticed that the hobbitd.pid file did not get deleted when
hobbitd terminated.  It seems like it should have been.

Many thanks,

Sebastian
--
Josh Luthman
Office: XXX-XXX-XXXX
Direct: XXX-XXX-XXXX
XXXX Wayne St
Suite XXXX
Troy, OH XXXXX

Those who don't understand UNIX are condemned to reinvent it, poorly.
--- Henry Spencer

-- 
Josh Luthman
Office: XXX-XXX-XXXX
Direct: XXX-XXX-XXXX
XXXX Wayne St
Suite XXXX
Troy, OH XXXXX

Those who don't understand UNIX are condemned to reinvent it, poorly.
--- Henry Spencer